Classification and Clustering Based Ensemble Techniques for Intrusion Detection Systems: A Survey

Abstract

Abstract A huge amount of data is transmitted through the networks, which allowed the exchange of knowledge and medical expertise, trade and banking facilities, etc. However, due to the huge connections to these networks, the security issue has been floated on the surface. Intrusion Detection System (IDS) plays a significant role to protect computer systems. To compensate these issues, the orientation is to employed machine learning and data mining techniques to design and implement powerful IDSs. Among these techniques is ensemble learning which enables a combination of multiple models to enhance overall performance. This study presents a brief overview of IDSs, discusses the history of ensemble systems, specifies the methods adapted in designed such system, highlights the most important ensemble techniques, demonstrates in detail the main methods that have been adapted in combining ensemble components. Besides, special attention was paid to studies in the period (2009-2020) that focus onto both ensemble classification and clustering when developing IDSs.